Abstract
Purpose
Atopic dermatitis (AD) is a chronic inflammatory skin disorder with a multifactorial pathophysiology. Although AD has been characterized by a T helper type 2 cell response, the role of the myeloid populations in the pathogenesis of AD remains unclear.
Methods
Peripheral blood mononuclear cells from 48 AD patients and 48 healthy controls were profiled using mass cytometry, primarily focusing on dendritic cells (DCs) and monocytes. Further analysis of a public single-cell RNA sequencing (scRNA-seq) dataset and immunofluorescence staining of lesional skin in AD were conducted for further validation.
Results
The frequency of circulating cDC1 was significantly decreased in AD compared with healthy controls. The frequency of cDC1 was negatively correlated with disease severity scores and serum immunoglobulin E levels. The expression of FcεRIa was significantly increased in the DC populations, including cDC1, cDC2, plasmacytoid DC, and Axl+ DC. CD163, a marker of the inflammatory DC subset DC3, was increased in AD patients, suggesting an increased DC3 signature in AD patients. Analysis of a public scRNA-seq dataset further corroborated the decreased frequency of cDC1. The expression of cutaneous lymphocyte antigen was increased in cDC1 of AD compared with HC, suggesting increased migration of cDC1 to the skin. Aligned with this hypothesis, the frequency of cDC1 was shown to be increased in AD lesional skin using immunofluorescence staining.
Conclusions
These results provide insight into the potential role of DC and monocyte populations in AD. We report decreased circulating cDC1 frequency and increased DC3 signature. The corresponding increased frequency of cDC1 in AD lesional skin implies their role in modulating AD pathophysiology.
Keywords: Atopic dermatitis; dendritic cells; immune monitoring, monocytes
INTRODUCTION
Atopic dermatitis (AD) or atopic eczema is a chronic, relapsing inflammatory skin disorder with a lifetime prevalence of 10%–20%.1,2 Clinically, AD is characterized by recurrent eczematous lesions and pruritus, leading to discomfort and decreased quality of life. AD is a heterogeneous disease, exhibiting variability in age of onset, clinical presentation and disease course. In addition, AD is associated with increased risk of other atopic diseases including asthma and allergic rhinitis, as well as other inflammatory diseases such as arthritis.2,3 The pathophysiology of AD is multifactorial, including barrier dysfunction, dysbiosis, immune dysregulation, and neuroinflammation.4
Immunologically, AD patients exhibit a skewed T helper type 2 response. Activated T cells release Th2 type cytokines including interleukin (IL)-4 and IL-13, into the skin. The cytokines induce inflammation and activate downstream JAK pathways.5 In addition, Th2 cells induce antibody class switching to immunoglobulin E (IgE).6 Binding of IgE to FcεRI, the high-affinity IgE receptor expressed on basophils and mast cells, and crosslinking of IgE induce activation and degranulation of several inflammatory mediators.7 While most therapeutic strategies have focused on T cell-mediated pathophysiology and pathways important for T cell function, less attention has been given to other immune cells. In addition, as AD is heterogeneous, and the response rate to current treatment strategies varies among individuals, other approaches that could increase therapeutic efficacy are needed.
Although AD has been characterized by a Th2 response, the role of myeloid populations in the pathogenesis of AD remains unclear. Dendritic cells (DCs) also express FcεRI, and cell-bound IgE may take part in capture and internalization of allergens into DCs for antigen processing and presentation.8 In AD, inflammatory dendritic epidermal cells and dermal DCs have been suggested to induce type 2 immune responses.4 The elevated IgE levels, increased number of circulating eosinophils, and skin sensitization leading to systemic immune response and atopic march suggest that AD is a systemic disease and highlight the importance of profiling circulating immune cells in AD.9,10,11 Many studies characterizing immune cells in AD have focused on Th2 effector cells and, to some extent, skin-resident myeloid cell populations. However, the landscape of circulating DCs and monocytes is yet to be investigated. Moreover, recent studies have demonstrated the novel inflammatory circulating CD163+ monocyte-like DC populations, DC3, specifically in inflammatory disorders.12,13,14
Here, we used cytometry by time-of-flight (CyTOF) or mass cytometry to profile circulating immune cells of 48 AD patients compared to 48 healthy controls. We designed a panel focusing on DCs and monocytes for single-cell, high-dimensional immune profiling of circulating DC and monocyte populations in order to identify differences in their immune phenotypes.
MATERIALS AND METHODS
Subject information
A total of 48 adult AD patients and 48 healthy controls were recruited for this study (Fig. 1A). This study was approved by the Institutional Review Board (IRB) of Seoul National University Hospital (IRB No. 1908-031-1052 and 2108-123-1246) and the IRB of Eunpyeong St. Mary’s Hospital at the Catholic University of Korea (IRB No. PC22SISI0076), and all subjects provided written informed consent. Subject information is summarized in Table 1. Baseline characteristics are described by means and ranges for continuous variables. Category values are indicated by percentages. Variables were compared using the chi-square test and t-test for category and continuous variables, respectively. There was no statistically significant difference in the age or sex between groups. AD patients were stratified according to disease severity scores as measured by the eczema area and severity index (EASI) scores: mild (EASI < 16, n = 9), moderate (16 ≤ EASI < 23, n = 10), or severe (EASI ≥ 23, n = 26). Three AD patients whose EASI score was not measured were excluded from the disease severity analysis. The mean EASI score of the patients was 22.8. The median total IgE from blood test results of AD patients are shown in Table 1.
Fig. 1. Characterization of major immune cell populations in AD patients compared with healthy controls. (A) Experimental workflow. PBMCs were isolated from 48 AD patients and 48 healthy controls. After staining, samples were acquired on the Fluidigm Helios mass cytometer. Data were analyzed using dimensionality reduction, phenotypic characterization and correlation analyses. (B) viSNE plots of 100,000 events per group overlaid with major immune cell populations. (C) Heatmap showing median marker expression patterns. (D) Frequencies of the myeloid cell populations represented as a percentage of CD45+ cells. Statistical analysis performed using unpaired t-test with Welch’s correction. Blue, HC; orange, AD.
HC, healthy control; AD, atopic dermatitis; PBMC, peripheral blood mononuclear cell; CyTOF, cytometry by time-of-flight; DC, dendritic cell; pDC, plasmacytoid DC.
An asterisk (*) denotes significant changes. **P < 0.01, ***P < 0.001.
Table 1. Demographic characteristics of study participants.
| Characteristics | AD (n = 48) | Control (n = 48) | P value | |
|---|---|---|---|---|
| Age (yr) | 32.3 (19–60) | 32.8 (21–56) | 0.815 | |
| Sex | 0.083 | |||
| Female | 12 (25.0) | 20 (41.7) | ||
| Male | 36 (75.0) | 28 (58.3) | ||
| AD severity | ||||
| Severe | 26 (54.2) | |||
| Moderate | 10 (20.8) | |||
| Mild | 9 (18.8) | |||
| No score | 3 (6.3) | |||
| Total IgE (IU/mL) | 1,383 (59.1–161,000) | |||
Values are presented as number (%) or median (range).
AD, atopic dermatitis; IgE, immunoglobulin E.
Sample processing
The experimental scheme with CyTOF analysis is illustrated in Fig. 1A. Peripheral blood was collected in sodium heparin Vacutainer Mononuclear Cell Preparation Tubes (BD Biosciences, Franklin Lakes, NJ, USA). Peripheral blood mononuclear cells (PBMCs) were isolated from the blood of healthy controls and AD patients by gradient centrifugation at 1,800 g for 20 minutes at room temperature. After washing cells with RPMI medium containing 10% fetal bovine serum, PBMCs were counted and cryopreserved until use.
CyTOF staining and acquisition
PBMCs were thawed and washed with cell medium. Cells were incubated in Rh103 Intercalator at 37°C for 20 minutes to label dead cells. After blocking with Fc receptor blocking solution to prevent nonspecific binding, cells were stained with a cocktail of commercial antibodies from Fluidigm or in-house conjugated antibodies listed in Supplementary Table S1 using Fluidigm’s Maxpar X8 polymer kits on ice for 30 minutes. Cells were then barcoded using Fluidigm’s Cell-ID 20-Plex Pd Barcoding Kit. EQ beads were spiked to calibrate and normalize between batches. The samples were acquired on Fluidigm Helios Mass Cytometer at an event rate of 400 cells/s. Acquired data were preprocessed to normalize, concatenate, debarcode, and remove EQ beads.
CyTOF data analysis
Data analysis was conducted using Cytobank. Gating strategy is shown in Supplementary Fig. S1. After preprocessing for cells to exclude residual beads, DNA level and Gaussian parameters were used to discriminate debris and doublets.15 Then, dead cells were excluded based on Rh103 expression. CD45+CD66b- cells were gated for downstream annotation. Data were analyzed using a supervised, manual approach using known markers for immune cell subsets as well as a semi-supervised clustering approach using FlowSOM.16 For supervised analysis, broad immune cell types were manually annotated using cell lineage markers. The frequency of each annotated population as a percentage of total CD45+ immune cells for each patient was exported to determine differences between healthy control and AD patients. To characterize marker expression profiles, the marker expression for each annotated cell population was exported as median signal intensity (MSI) values. Arcsinh-transformed MSI values were used for analysis. Heatmaps were drawn with Clustergrammer,17 using average linkage and cosine distance. To test correlation between cell frequency and disease severity or serum IgE levels, Pearson correlation analyses were conducted. To complement our manual analysis using canonical markers and further investigate changes in the DC and monocyte populations, semi-supervised clustering of these cells was conducted using FlowSOM, which builds a tree using self-organizing map (SOM) clustering to compute metaclustering of mass cytometry data. The metaclusters’ frequencies out of myeloid cells and median marker expression were exported to characterize myeloid cells.
Analysis of public single-cell RNA sequencing dataset
To confirm our CyTOF results, scRNA-seq data obtained from healthy control (n = 6) and AD (n = 12) PBMCs were analyzed.18 scRNA-seq data were aligned to the human reference genome using Cellranger v7.1.0. For preprocessing and quality control, cells were filtered using a mitochondrial gene percentage threshold of under ten percent and gene counts between 200 and 5,000. Doublets were removed using scDblFinder. Canonical correlation analysis integration was conducted to remove batch effects. Cells were annotated into broad immune cell subsets using lineage markers. DCs and monocytes were subclustered and further annotated into CD14+ monocyte, CD14+CD16+ monocyte, CD16+ monocyte, cDC1, cDC2, plasmacytoid DC (pDC) and Axl+ DC. The frequency of each annotated population as a percentage of total DCs and monocytes was calculated. The differential expression of genes related to skin homing was evaluated between HC and AD.
Immunofluorescence (IF) staining
Skin samples were obtained from non-AD control patients (n = 6) and AD patients (n = 8). For histological examination, the specimens were fixed in 10% formalin, embedded in paraffin, cut into 5 µm sections, and attached to silane coated slides. Slides were deparaffinized with xylene and ethanol and then boiled in antigen retrieval buffer (Novus Biologicals, Littleton, CO, Canada) using a microwave. After blocking with SuperBlock buffer (Thermo Scientific, Waltham, MA, USA), slides were immunostained with an anti-CLEC9A antibody (1:100, Abcam, Cambridge, UK) overnight at 4°C. Slides were washed three times with PBS and incubated with anti-rabbit AlexaFluor-594 secondary antibody (1:200, Abcam) at room temperature for 2 hours, rinsed three times with PBS and stained with FITC conjugated anti-CD1c antibody (1:100, Invitrogen, Waltham, MA, USA) and AlexaFluor-647 conjugated anti-CD11c antibody (1:100, Cell Signaling, Danvers, MA, USA) for 2 hours. The slides were stained with DAPI. Quantification was performed by counting the number of positive cells under 200× magnification.
Statistical analysis
Statistical analyses were performed using GraphPad Prism v9 (GraphPad Software Inc., San Diego, CA, USA). To compare the cell frequencies and the median marker expression between healthy control and AD, unpaired t-test with Welch correction was used. For comparison of cell frequency among disease severity groups, Brown-Forsythe and Welch ANOVA tests were used. For IF quantification, unpaired t-test with Welch correction was used. P values < 0.05 were considered statistically significant.
RESULTS
Changes in the myeloid compartment in AD patients
To identify changes in the immune cell composition in the PBMCs of AD patients compared with healthy controls, major immune cell populations were first annotated using canonical cell lineage markers (Supplementary Fig. S1). viSNE plots of 100,000 events for healthy control and AD groups with the annotated populations showed 14 major immune cell types (Fig. 1B) and the respective marker expression patterns (Fig. 1C). In AD, among the myeloid cell compartments, CD14+CD16lo monocyte frequency was increased, while cDC1 and pDC were significantly decreased (Fig. 1D). The frequencies of other immune cell types are shown in Supplementary Fig. S2. CD4+ T cell frequency was elevated in AD. Aligned with previous reports, the natural killer (NK) cell population was significantly decreased in AD (Supplementary Fig. S2).19
Phenotypic characterization of myeloid cells between AD and healthy controls
Next, to further investigate the DC compartment in AD, we evaluated the differences in marker expression level. Fig. 2A shows volcano plots demonstrating differentially expressed markers in AD versus healthy controls. FcεRIa, the high affinity IgE receptor, was significantly elevated in cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+ DC and pDC populations of AD patients (Fig. 2B-F). The increase in FcεRIα expression suggests a response to increased IgE levels in AD patients. In addition, CLEC9A, a marker of cDC1, was decreased in the cDC1 population of AD patients (Fig. 2B). HLA-DR expression was decreased in cDC2(CD14lo), cDC2(CD14hi) and Axl+ DC (Fig. 2C, D, and E). The expression of CD163, which is a marker for the inflammatory DC subset DC3, was increased in cDC2(CD14lo) in AD, suggesting an increased DC3 signature in AD. DC3 has recently been classified as an inflammatory subset of DCs.14 Similarly, differentially expressed markers in AD versus healthy controls for the monocyte populations are shown in Supplementary Fig. S3A. CD11b and HLA-DR was decreased in CD14loCD16+ monocytes (Supplementary Fig. S3B). The expression of PD-L1 was increased in CD14loCD16+ and CD14+CD16+ monocytes in AD, and CD40 was increased in CD14+CD16+ monocytes (Supplementary Fig. S3B and C).
Fig. 2. Differentially expressed functional markers of dendritic cells between AD patients and healthy controls. (A) Volcano plots showing differential marker expression between HC and AD. Red, increased expression in AD; blue, decreased expression in AD. (B-E) Violin plots showing significantly different marker expression in (B) cDC1, (C) cDC2(CD14lo), (D) cDC2(CD14hi), (E) Axl+ DC, (F) pDC of AD patients compared with healthy controls. Statistical analysis performed using unpaired t-test with Welch’s correction.
HC, healthy control; AD, atopic dermatitis; DC, dendritic cell; pDC, plasmacytoid DC.
An asterisk (*) denotes significant changes. **P < 0.01, ***P < 0.001, ****P < 0.0001.
The frequency of cDC1 further decreases in severe AD and is negatively correlated with disease severity and serum IgE levels
We stratified the AD patients into mild, moderate, or severe by disease severity, as measured by EASI scores. The viSNE plots of 25,000 events for each group with the annotated immune cell populations overlaid are shown (Fig. 3A). The frequency of cDC1 further decreased in more severe AD compared to moderate AD (Fig. 3B). Of the DC subsets, cDC1 was the only cell type that showed a gradual frequency change according to disease severity, as the frequencies of other cell types were not significantly altered among AD patients (Fig. 3B, Supplementary Fig. S4). Next, we conducted correlation analyses of the DC populations with disease severity scores and serum IgE levels. There was a weak negative correlation between cDC1 frequency and EASI score (r = −0.3023, P = 0.0436), and between cDC1 frequency and serum IgE levels (r = −0.2695, P = 0.1357) (Fig. 3C and D).
Fig. 3. Correlation analyses of the myeloid populations comparing healthy controls to mild, moderate, or severe AD patients. AD patients were stratified by disease severity into mild (n = 9), moderate (n = 10), or severe (n = 26), according to EASI scores. (A) viSNE plots of 25,000 events per group overlaid with major immune cell populations. (B) Frequencies of DC populations (cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+ DC, and pDC) as a percentage of CD45+ cells according to disease severity. Statistical analysis performed using Brown-Forsythe and Welch ANOVA test. (C) Pearson correlation of cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+ DC, or pDC frequency and EASI score. (D) Pearson correlation of cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+ DC, or pDC frequency with serum log IgE levels.
HC, healthy control; DC, dendritic cell; pDC, plasmacytoid DC; IgE, immunoglobulin E; EASI, eczema area and severity index; AD, atopic dermatitis.
An asterisk (*) denotes significant changes. *P < 0.05, **P < 0.01, ****P < 0.0001.
Semi-supervised analysis reveals a decrease in cDC1 frequency and an increased DC3 signature
Next, we performed semi-supervised clustering of myeloid cells to confirm our manual findings. FlowSOM clustering uses a self-organizing map for clustering and builds a tree by computing metaclusters, allowing for the detection of subsets that were potentially missed with a manual approach.16 FlowSOM analysis of the myeloid populations identified ten metaclusters (Fig. 4A). The metaclusters were annotated based on the expression patterns shown in Fig. 4B. cDC1, two subsets of cDC2(cDC2(CD14lo) and cDC2(CD14hi)), two subsets of Axl+ DC(Axl+CD169+ DC and Axl+HLADRlo DC), pDC, CD1c-CD141- DC and three monocyte populations were annotated (Fig. 4C). Importantly, the frequency of cDC1 (metacluster 7) also significantly decreased in AD patients with this analysis approach, confirming our supervised analysis (Fig. 4C). Next, we evaluated the phenotypic markers of DCs (Fig. 5) and monocytes (Supplementary Fig. S5) comparing healthy control and AD. Volcano plots show the differential marker expression in cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+CD169+ DC, pDC, and CD1c-CD141- DC (Fig. 5A). In addition, we validated our manual findings and further characterized the phenotypic differences in the DC populations using this approach. Markers that were significantly altered in AD compared to controls were highlighted in Fig. 5B-G. FcεRIα expression was increased in cDC1, cDC2(CD14lo), cDC2(CD14hi), Axl+CD169+DC, and pDC (Fig. 5B-F). CLEC9A was decreased in cDC1 of AD patients (Fig. 5B). Similar to our manual approach, CD163 was increased in both populations of cDC2 of AD patients (Fig. 5C and D), implicating an increased DC3 signature in AD. HLA-DR was decreased in cDC2(CD14lo), and cDC2(CD14hi), while it was increased in CD1c-CD141- DC (Fig. 5C, D, and G). In addition, CD40 and CD38 were decreased in CD1c-CD141- DC (Fig. 5G). Collectively, our results suggest a decreased cDC1 frequency and an increased DC3 signature in AD.
Fig. 4. Semi-supervised clustering analysis of myeloid cells. (A) FlowSOM tree resulting from semi-supervised clustering of the myeloid cell population identified ten metaclusters. (B) Heatmap showing median marker expression patterns of metaclusters. Metaclusters were annotated using these marker expressions. (C) Frequencies of metaclusters as a percentage of total myeloid cells. Statistical analysis performed using unpaired t-test with Welch’s correction. Blue, HC; orange, AD.
MC, metacluster; HC, healthy control; AD, atopic dermatitis; DC, dendritic cell; pDC, plasmacytoid DC.
An asterisk (*) denotes significant changes. **P < 0.01, ***P < 0.001, ****P < 0.0001.
Fig. 5. Differential marker expression analysis of dendritic cell metaclusters between AD patients and healthy controls. (A) Volcano plots showing differential marker expression between HC and AD. Red, increased expression in AD; blue, decreased expression in AD. (B-E) Violin plots showing significantly different marker expression in (B) MC7 cDC1, (C) MC2 cDC2(CD14lo), (D) MC4 pDC, (E) MC3 Axl+ DC, (F) MC6 cDC2(CD14hi) and (G) MC5 CD1c-CD141- DC of AD patients compared with healthy controls. Statistical analysis performed using unpaired t-test with Welch’s correction.
DC, dendritic cell; pDC, plasmacytoid DC; HC, healthy control; AD, atopic dermatitis; MC, metacluster.
An asterisk (*) denotes significant changes. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Analysis of scRNA-seq data shows decreased frequency of circulating cDC1 in AD patients
To confirm our finding of decreased cDC1 frequency and to further explore the characteristics of this subset, we analyzed a published scRNA-seq dataset of PBMCs from AD patients and HC.18 After preprocessing and QC, immune cells (HC, 98,688; AD, 186,261 cells) were annotated using canonical lineage markers (Fig. 6A and B). DCs and monocytes were subclustered and annotated into specific cell populations depending on marker expression as shown in Fig. 6C and D. Similar to our CyTOF results, the frequency of cDC1 was significantly decreased in AD (Fig. 6E). The expression of CD163 was increased, suggesting elevated DC3 signature (Fig. 6F). In addition, the expression of SELPLG, the gene encoding cutaneous lymphocyte antigen, a skin-homing marker, was significantly increased in the cDC1 population of AD compared to HC, suggesting a higher skin-homing potential of cDC1 to AD lesional skin (Fig. 6G). Moreover, the expression of skin-homing chemokine receptors was increased in cDC1 (Fig. 6H). These results validate our finding of decreased cDC1 frequency and increased DC3 signature, suggesting that cDC1 have increased potential to migrate to the skin.
Fig. 6. scRNA-seq data shows decreased frequency of circulating cDC1 and increased skin-homing receptors in cDC1 of AD patients. (A) UMAP visualization of 284,949 cells (HC, 98,688 cells; AD, 186,261 cells). (B) Clusters were annotated into broad immune cell subsets based on the expression of lineage markers. (C) UMAP visualization of monocytes and DCs (total, 60,365 cells; HC, 20,503 cells; AD, 39,862 cells). (D) Clusters were further annotated based on the expression of select markers. (E) Violin plots frequency of each cluster as a percentage of total monocyte/DC populations. (F) Violin plot showing the expression of CD163 in cDC2(CD14hi). (G) Violin plot showing the expression of SELPLG in cDC1, which encodes cutaneous lymphocyte antigen, cutaneous lymphocyte antigen (H) Dotplot showing the differential expression of chemokine receptors CCR2, CCR4, CCR5 and CCR6 between cDC1 of HC and AD. Statistical analysis performed using unpaired t-test with Welch’s correction.
DC, dendritic cell; pDC, plasmacytoid DC; UMAP, uniform manifold approximation and projection; HC, healthy control; AD, atopic dermatitis.
An asterisk (*) denotes significant changes. *P < 0.05, **P < 0.01, ****P < 0.0001.
Immunofluorescence staining shows increased cDC1 in the lesional skin of AD patients
To confirm our hypothesis that the decreased cDC1 frequency in the circulation in AD patients may have been due to its increased recruitment to the lesional skin, we obtained skin biopsies from AD lesions (n = 8) and non-AD skin (n = 6) and performed IF staining with CLEC9A, CD1c and CD11c. Representative images are shown in Fig. 7A. Compared to non-AD skin, the proportion of cDC1 among CD11c+ DCs was significantly increased in AD lesional skin (Fig. 7B). The proportion of cDC2 among CD11c+ DCs was similar between control skin and AD lesional skin (Supplementary Fig. S6). Therefore, our findings demonstrate that the frequency of cDC1 is significantly decreased in the circulation of AD patients, both by CyTOF and scRNA-seq analyses, and show that the proportion of cDC1 among total DCs is increased in the lesional skin.
Fig. 7. Immunofluorescence staining showing increased cDC1 frequency in AD lesional skin. (A) Representative images (200× magnification) showing CLEC9A and CD11c expression in skin of control and AD skin. Scale bar indicates 20 μm. (B) Immunofluorescence staining results were quantified using the number of positive cells per 200x field, and the frequency of cDC1 (CLEC9A+CD11c+ cells as a total of CD11c+ cells) was calculated. CTRL, n = 6; AD, n = 8.
CTRL, control; AD, atopic dermatitis; DC, dendritic cell.
An asterisk (*) denotes significant changes. *P < 0.05.
DISCUSSION
DCs, as discovered by Ralph Steinman and Zanvil Cohn, are mononuclear phagocytes in blood and diverse tissues that ingest and present antigens to T cells and regulate innate and adaptive immune responses.20,21 In addition to CD141+CLEC9A+ cDC1, CD1c+CLEC10A+ cDC2 and CD123+ pDC, recent studies using single-cell RNA sequencing have allowed for the discovery of novel subsets including CD5-CD163+ DC3, CD141-CD1c- DC4 and AXL+ DC5.14 Of these additional subsets, DC3 is a subpopulation of cDC2 and was shown to be expanded in inflammatory contexts.12,22 The proinflammatory cytokines or chemokines released by activated DCs may in turn contribute to AD pathophysiology. Conversely, immune regulatory DCs may play roles in T cell tolerance by producing IL-10 and promoting the expansion of regulatory T cells.23 The exact role and function of DC subsets including these novel subsets especially in AD contexts remain largely unknown.
Here, we report a decrease in the frequency of cDC1 in the blood of AD patients, as analyzed by mass cytometry as well as scRNA-seq. The frequency of cDC1 is low in the circulation and tissue, and its role in AD remains largely unexplored. cDC1 has been described in the literature for its role in cross-presentation of dead cells to CD8+ T cells.24 Changes in its frequency have been demonstrated in some diseases, such as allergic rhinitis and systemic lupus erythematosus.25,26 In inflammatory arthritis, the frequency of cDC1 is increased in synovial fluid, but these cDC1s have a quiescent phenotype, without a clear inflammatory profile.27 Indeed, cDC1 expresses many immunoregulatory molecules, and its regulatory or tolerogenic roles have been reported in several contexts.24 For instance, in mice, a subset of cDC1 has been reported to induce regulatory T cells.28 In addition, cDC1 in human skin has been reported to produce IL-10 and induce regulatory T cells that suppress inflammation.29 We hypothesized that the decreased frequency of circulating cDC1 in AD was due to its increased migration to the lesional skin, and supported this by showing that the frequency of cDC1 among total DCs was increased in the skin of AD patients through IF straining. Although the exact role of cDC1 in AD pathophysiology remains to be explored, the increased infiltration in AD lesional skin, and the negative correlation between circulating cDC1 frequency and disease severity scores suggest that cDC1 may have migrated to the lesional skin as a response to inflammation and increased Th2 responses as a tolerogenic mechanism.
The expression of CD163 was increased in AD patients compared to healthy controls, suggesting an increased DC3 signature. DC3, distinguished by CD5-CD163+, is considered to be an inflammatory DC subset originating from cDC2. DC3 has been shown to be increased in the peripheral blood of patients with systemic lupus erythematosus and coronavirus disease 2019 patients, and has been linked with inflammation.12,30 In addition to inflammation, DC3 has been reported to promote Th17 polarization.12 The Asian AD phenotype displays increased Th17 polarization, with a higher induction of Th17-related cytokines compared with European American AD.31 In addition, the number of Th17 cells was increased in the peripheral blood of AD patients32 compared to healthy controls. Further exploration is necessary to investigate whether the increase in DC3 signature observed in Korean AD populations is a unique feature of Asian AD or is a general feature of AD.
Our current findings support the previous studies showing differences between healthy control and AD patients. For instance, FcεRI, which was increased on several DC populations in this study, has been reported to be increased in AD.33 Although many studies have reported increased FcεRI in AD, whether FcεRI upregulation precedes or results from IgE elevation remains unclear.8,34,35,36 Allergens that bind IgE/FcεRI complexes are internalized and loaded onto MHC class II for presentation to T cells.8 In addition, the elevated serum IgE in AD could in turn sustain surface expression of FcεRI on DCs.37 Moreover, IL-4 has been reported to enhance FcεRI expression on human monocytes and DCs.37,38,39,40 Therefore, the global elevation of FcεRI in DCs may play an important role in AD pathogenesis via allergen capture, internalization and Th2 polarization.41 In addition, we noted that the frequency of circulating NK cells was significantly decreased in AD patients. The frequency of NK cells has been reported to be decreased in the circulation of AD patients, and strategies aiming to restore NK cells have been suggested to be a possible immunotherapeutic strategy.19 We found that the frequency of pDC was decreased in our cohort. There have been reports that pDC is decreased in AD lesion, which may be correlated with higher susceptibility to cutaneous infection in the lesions.42 In the circulation, there have been mixed reports regarding the frequency of pDC. A previous report demonstrated increased pDC frequency compared to healthy controls, and lower myeloid DC to pDC ratio compared to controls.43 Concurrently, however, others have reported that there are minimal changes in the frequency of pDC in AD compared to healthy individuals.44 Although not directly in AD patients, one study has reported that upon allergen challenge, blood pDCs are rapidly reduced and then return to baseline after 24 hours in atopic individuals, suggesting the pDC frequency may dynamically change depending on disease activity.45 These variable findings may be due to disease heterogeneity, age, ethnicity or profiling method. However, we found that the expression of FcεRIa is increased in pDCs in AD compared to healthy controls, which is aligned with previous reports.43,44 Even with reduced pDC numbers, their heightened heightened FcεRIa expression may amplify IgE-mediated signaling to exacerbate Th2 inflammation. In-depth studies exploring specific subsets are warranted.
Some limitations of this study include that the majority of the cohort involved chronic, severe patients of the same ethnicity. Further studies on additional cohorts may be needed to generalize our results. In addition, our study is limited in its descriptive nature, with additional research required to investigate whether there are changes in the frequency and phenotype of several DC subsets in the lesional skin in conjunction with global changes. Furthermore, the role of the alterations in DCs, especially the decrease in circulating cDC1 and increased migration to the skin in the pathophysiology of AD remains to be investigated. We are also planning to further investigate these cell subsets in AD lesional skin using spatial transcriptomic or proteomic analysis. This study is the first to explore the myeloid populations at the single-cell level in AD using mass cytometry in a large cohort of 48 AD patients and 48 healthy controls. We validated the decreased circulating cDC1 frequency using a public scRNA-seq dataset and showed increased cDC1 frequency in AD lesional skin through IF staining. Thus, we have conducted a high-dimensional profiling of circulating DCs and monocytes in AD patients and provide insight in the potential role of these cell populations in AD pathophysiology.
ACKNOWLEDGMENTS
This research was supported by the Bio & Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. RS-2022-NR067309). This research was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science (2020R1F1A1073692). This research was funded by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (2019R1F1A1060787 and 2022R1A2C1010554).
Footnotes
Disclosure: There are no financial or other issues that might lead to conflict of interest.
Data Availability Statement: The data that support the findings of this study are available from the corresponding author upon reasonable request. Some data are not publicly available due to privacy or ethical restrictions. Single-cell transcriptomic data were obtained from Jin et al.18
SUPPLEMENTARY MATERIALS
CyTOF antibody panel
Gating strategy.
Characterization of non-myeloid cell populations in AD patients compared with age-, sex-matched healthy controls.
Differentially expressed functional markers of monocytes between AD patients and healthy controls.
Frequencies of other immune cell populations comparing healthy controls to mild, moderate, or severe AD patients.
Differential marker expression analysis of monocyte metaclusters between AD patients and healthy controls.
Immunofluorescence staining showing no change in cDC2 frequency in AD lesional skin.
References
- 1.Deckers IA, McLean S, Linssen S, Mommers M, van Schayck CP, Sheikh A. Investigating international time trends in the incidence and prevalence of atopic eczema 1990-2010: a systematic review of epidemiological studies. PLoS One. 2012;7:e39803. doi: 10.1371/journal.pone.0039803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Silverberg JI, Hanifin JM. Adult eczema prevalence and associations with asthma and other health and demographic factors: a US population-based study. J Allergy Clin Immunol. 2013;132:1132–1138. doi: 10.1016/j.jaci.2013.08.031. [DOI] [PubMed] [Google Scholar]
- 3.Schmitt J, Schwarz K, Baurecht H, Hotze M, Fölster-Holst R, Rodríguez E, et al. Atopic dermatitis is associated with an increased risk for rheumatoid arthritis and inflammatory bowel disease, and a decreased risk for type 1 diabetes. J Allergy Clin Immunol. 2016;137:130–136. doi: 10.1016/j.jaci.2015.06.029. [DOI] [PubMed] [Google Scholar]
- 4.Weidinger S, Beck LA, Bieber T, Kabashima K, Irvine AD. Atopic dermatitis. Nat Rev Dis Primers. 2018;4:1. doi: 10.1038/s41572-018-0001-z. [DOI] [PubMed] [Google Scholar]
- 5.Bhattacharjee A, Shukla M, Yakubenko VP, Mulya A, Kundu S, Cathcart MK. IL-4 and IL-13 employ discrete signaling pathways for target gene expression in alternatively activated monocytes/macrophages. Free Radic Biol Med. 2013;54:1–16. doi: 10.1016/j.freeradbiomed.2012.10.553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Brandt EB, Sivaprasad U. Th2 cytokines and atopic dermatitis. J Clin Cell Immunol. 2011;2:110. doi: 10.4172/2155-9899.1000110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kinet JP. The high-affinity IgE receptor (Fc epsilon RI): from physiology to pathology. Annu Rev Immunol. 1999;17:931–972. doi: 10.1146/annurev.immunol.17.1.931. [DOI] [PubMed] [Google Scholar]
- 8.Maurer D, Fiebiger E, Reininger B, Ebner C, Petzelbauer P, Shi GP, et al. Fc epsilon receptor I on dendritic cells delivers IgE-bound multivalent antigens into a cathepsin S-dependent pathway of MHC class II presentation. J Immunol. 1998;161:2731–2739. [PubMed] [Google Scholar]
- 9.Bantz SK, Zhu Z, Zheng T. The atopic march: progression from atopic dermatitis to allergic rhinitis and asthma. J Clin Cell Immunol. 2014;5:202. doi: 10.4172/2155-9899.1000202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Darlenski R, Kazandjieva J, Hristakieva E, Fluhr JW. Atopic dermatitis as a systemic disease. Clin Dermatol. 2014;32:409–413. doi: 10.1016/j.clindermatol.2013.11.007. [DOI] [PubMed] [Google Scholar]
- 11.Thijs JL, Strickland I, Bruijnzeel-Koomen CAFM, Nierkens S, Giovannone B, Knol EF, et al. Serum biomarker profiles suggest that atopic dermatitis is a systemic disease. J Allergy Clin Immunol. 2018;141:1523–1526. doi: 10.1016/j.jaci.2017.12.991. [DOI] [PubMed] [Google Scholar]
- 12.Dutertre CA, Becht E, Irac SE, Khalilnezhad A, Narang V, Khalilnezhad S, et al. Single-cell analysis of human mononuclear phagocytes reveals subset-defining markers and identifies circulating inflammatory dendritic cells. Immunity. 2019;51:573–589.e8. doi: 10.1016/j.immuni.2019.08.008. [DOI] [PubMed] [Google Scholar]
- 13.Ginhoux F, Guilliams M, Merad M. Expanding dendritic cell nomenclature in the single-cell era. Nat Rev Immunol. 2022;22:67–68. doi: 10.1038/s41577-022-00675-7. [DOI] [PubMed] [Google Scholar]
- 14.Villani AC, Satija R, Reynolds G, Sarkizova S, Shekhar K, Fletcher J, et al. Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors. Science. 2017;356:eaah4573. doi: 10.1126/science.aah4573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lee BH, Rahman AH. Acquisition, processing, and quality control of mass cytometry data. Methods Mol Biol. 2019;1989:13–31. doi: 10.1007/978-1-4939-9454-0_2. [DOI] [PubMed] [Google Scholar]
- 16.Van Gassen S, Callebaut B, Van Helden MJ, Lambrecht BN, Demeester P, Dhaene T, et al. FlowSOM: using self-organizing maps for visualization and interpretation of cytometry data. Cytometry A. 2015;87:636–645. doi: 10.1002/cyto.a.22625. [DOI] [PubMed] [Google Scholar]
- 17.Fernandez NF, Gundersen GW, Rahman A, Grimes ML, Rikova K, Hornbeck P, et al. Clustergrammer, a web-based heatmap visualization and analysis tool for high-dimensional biological data. Sci Data. 2017;4:170151. doi: 10.1038/sdata.2017.151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jin SP, Lee K, Bang YJ, Jeon YH, Jung S, Choi SJ, et al. Mapping the immune cell landscape of severe atopic dermatitis by single-cell RNA-seq. Allergy. 2024;79:1584–1597. doi: 10.1111/all.16121. [DOI] [PubMed] [Google Scholar]
- 19.Mack MR, Brestoff JR, Berrien-Elliott MM, Trier AM, Yang TB, McCullen M, et al. Blood natural killer cell deficiency reveals an immunotherapy strategy for atopic dermatitis. Sci Transl Med. 2020;12:eaay1005. doi: 10.1126/scitranslmed.aay1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Merad M, Sathe P, Helft J, Miller J, Mortha A. The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting. Annu Rev Immunol. 2013;31:563–604. doi: 10.1146/annurev-immunol-020711-074950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Steinman RM, Cohn ZA. Identification of a novel cell type in peripheral lymphoid organs of mice. I. Morphology, quantitation, tissue distribution. J Exp Med. 1973;137:1142–1162. doi: 10.1084/jem.137.5.1142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bourdely P, Anselmi G, Vaivode K, Ramos RN, Missolo-Koussou Y, Hidalgo S, et al. Transcriptional and functional analysis of CD1c+ human dendritic cells identifies a CD163+ subset priming CD8+CD103+ T cells. Immunity. 2020;53:335–352.e8. doi: 10.1016/j.immuni.2020.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Banerjee DK, Dhodapkar MV, Matayeva E, Steinman RM, Dhodapkar KM. Expansion of FOXP3high regulatory T cells by human dendritic cells (DCs) in vitro and after injection of cytokine-matured DCs in myeloma patients. Blood. 2006;108:2655–2661. doi: 10.1182/blood-2006-03-011353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Balan S, Radford KJ, Bhardwaj N. Unexplored horizons of cDC1 in immunity and tolerance. Adv Immunol. 2020;148:49–91. doi: 10.1016/bs.ai.2020.10.002. [DOI] [PubMed] [Google Scholar]
- 25.Lundberg K, Greiff L, Borrebaeck CA, Lindstedt M. FcεRI levels and frequencies of peripheral blood dendritic cell populations in allergic rhinitis. Hum Immunol. 2010;71:931–933. doi: 10.1016/j.humimm.2010.07.008. [DOI] [PubMed] [Google Scholar]
- 26.Migita K, Miyashita T, Maeda Y, Kimura H, Nakamura M, Yatsuhashi H, et al. Reduced blood BDCA-2+ (lymphoid) and CD11c+ (myeloid) dendritic cells in systemic lupus erythematosus. Clin Exp Immunol. 2005;142:84–91. doi: 10.1111/j.1365-2249.2005.02897.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Boltjes A, Samat AAK, Plantinga M, Mokry M, Castelijns B, Swart JF, et al. Conventional dendritic cells type 1 are strongly enriched, quiescent and relatively tolerogenic in local inflammatory arthritis. Front Immunol. 2023;13:1101999. doi: 10.3389/fimmu.2022.1101999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Guilliams M, Crozat K, Henri S, Tamoutounour S, Grenot P, Devilard E, et al. Skin-draining lymph nodes contain dermis-derived CD103(-) dendritic cells that constitutively produce retinoic acid and induce Foxp3(+) regulatory T cells. Blood. 2010;115:1958–1968. doi: 10.1182/blood-2009-09-245274. [DOI] [PubMed] [Google Scholar]
- 29.Chu CC, Ali N, Karagiannis P, Di Meglio P, Skowera A, Napolitano L, et al. Resident CD141 (BDCA3)+ dendritic cells in human skin produce IL-10 and induce regulatory T cells that suppress skin inflammation. J Exp Med. 2012;209:935–945. doi: 10.1084/jem.20112583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kvedaraite E, Hertwig L, Sinha I, Ponzetta A, Hed Myrberg I, Lourda M, et al. Major alterations in the mononuclear phagocyte landscape associated with COVID-19 severity. Proc Natl Acad Sci U S A. 2021;118:e2018587118. doi: 10.1073/pnas.2018587118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Noda S, Suárez-Fariñas M, Ungar B, Kim SJ, de Guzman Strong C, Xu H, et al. The Asian atopic dermatitis phenotype combines features of atopic dermatitis and psoriasis with increased TH17 polarization. J Allergy Clin Immunol. 2015;136:1254–1264. doi: 10.1016/j.jaci.2015.08.015. [DOI] [PubMed] [Google Scholar]
- 32.Koga C, Kabashima K, Shiraishi N, Kobayashi M, Tokura Y. Possible pathogenic role of Th17 cells for atopic dermatitis. J Invest Dermatol. 2008;128:2625–2630. doi: 10.1038/jid.2008.111. [DOI] [PubMed] [Google Scholar]
- 33.Beeren IM, De Bruin-Weller MS, Ra C, Kok I, Bruinzeel-Koomen CA, Knol EF. Expression of Fc(epsilon)RI on dendritic cell subsets in peripheral blood of patients with atopic dermatitis and allergic asthma. J Allergy Clin Immunol. 2005;116:228–229. doi: 10.1016/j.jaci.2005.02.039. [DOI] [PubMed] [Google Scholar]
- 34.Maurer D, Ebner C, Reininger B, Fiebiger E, Kraft D, Kinet JP, et al. The high affinity IgE receptor (Fc epsilon RI) mediates IgE-dependent allergen presentation. J Immunol. 1995;154:6285–6290. [PubMed] [Google Scholar]
- 35.Maurer D, Fiebiger S, Ebner C, Reininger B, Fischer GF, Wichlas S, et al. Peripheral blood dendritic cells express Fc epsilon RI as a complex composed of Fc epsilon RI alpha- and Fc epsilon RI gamma-chains and can use this receptor for IgE-mediated allergen presentation. J Immunol. 1996;157:607–616. [PubMed] [Google Scholar]
- 36.Novak N, Allam JP, Hagemann T, Jenneck C, Laffer S, Valenta R, et al. Characterization of FcepsilonRI-bearing CD123 blood dendritic cell antigen-2 plasmacytoid dendritic cells in atopic dermatitis. J Allergy Clin Immunol. 2004;114:364–370. doi: 10.1016/j.jaci.2004.05.038. [DOI] [PubMed] [Google Scholar]
- 37.Novak N, Tepel C, Koch S, Brix K, Bieber T, Kraft S. Evidence for a differential expression of the FcepsilonRIgamma chain in dendritic cells of atopic and nonatopic donors. J Clin Invest. 2003;111:1047–1056. doi: 10.1172/JCI15932. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Geiger E, Magerstaedt R, Wessendorf JH, Kraft S, Hanau D, Bieber T. IL-4 induces the intracellular expression of the alpha chain of the high-affinity receptor for IgE in in vitro-generated dendritic cells. J Allergy Clin Immunol. 2000;105:150–156. doi: 10.1016/s0091-6749(00)90190-8. [DOI] [PubMed] [Google Scholar]
- 39.Reischl IG, Dubois GR, Peiritsch S, Brown KS, Wheat L, Woisetschläger M, et al. Regulation of Fc epsilonRI expression on human monocytic cells by ligand and IL-4. Clin Exp Allergy. 2000;30:1033–1040. doi: 10.1046/j.1365-2222.2000.00859.x. [DOI] [PubMed] [Google Scholar]
- 40.Gosset P, Lamblin-Degros C, Tillie-Leblond I, Charbonnier AS, Joseph M, Wallaert B, et al. Modulation of high-affinity IgE receptor expression in blood monocytes: opposite effect of IL-4 and glucocorticoids. J Allergy Clin Immunol. 2001;107:114–122. doi: 10.1067/mai.2001.111126. [DOI] [PubMed] [Google Scholar]
- 41.Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, et al. Reciprocal control of T helper cell and dendritic cell differentiation. Science. 1999;283:1183–1186. doi: 10.1126/science.283.5405.1183. [DOI] [PubMed] [Google Scholar]
- 42.Wollenberg A, Wagner M, Günther S, Towarowski A, Tuma E, Moderer M, et al. Plasmacytoid dendritic cells: a new cutaneous dendritic cell subset with distinct role in inflammatory skin diseases. J Invest Dermatol. 2002;119:1096–1102. doi: 10.1046/j.1523-1747.2002.19515.x. [DOI] [PubMed] [Google Scholar]
- 43.Hashizume H, Horibe T, Yagi H, Seo N, Takigawa M. Compartmental imbalance and aberrant immune function of blood CD123+ (plasmacytoid) and CD11c+ (myeloid) dendritic cells in atopic dermatitis. J Immunol. 2005;174:2396–2403. doi: 10.4049/jimmunol.174.4.2396. [DOI] [PubMed] [Google Scholar]
- 44.Stary G, Bangert C, Stingl G, Kopp T. Dendritic cells in atopic dermatitis: expression of FcepsilonRI on two distinct inflammation-associated subsets. Int Arch Allergy Immunol. 2005;138:278–290. doi: 10.1159/000088865. [DOI] [PubMed] [Google Scholar]
- 45.Vittorakis S, Samitas K, Tousa S, Zervas E, Aggelakopoulou M, Semitekolou M, et al. Circulating conventional and plasmacytoid dendritic cell subsets display distinct kinetics during in vivo repeated allergen skin challenges in atopic subjects. BioMed Res Int. 2014;2014:231036. doi: 10.1155/2014/231036. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
CyTOF antibody panel
Gating strategy.
Characterization of non-myeloid cell populations in AD patients compared with age-, sex-matched healthy controls.
Differentially expressed functional markers of monocytes between AD patients and healthy controls.
Frequencies of other immune cell populations comparing healthy controls to mild, moderate, or severe AD patients.
Differential marker expression analysis of monocyte metaclusters between AD patients and healthy controls.
Immunofluorescence staining showing no change in cDC2 frequency in AD lesional skin.